Lecturer: Qiu Yue

1. What is data analysis and why data analysis is needed

Obtain information and knowledge by sorting and processing data, so as to understand the survival of products, discover potential opportunities, guide and support product and operation decisions, and verify the effects of strategies.
Data is the gospel of Internet product managers, and it is also a core element that changes the nature of the work of Internet product managers and traditional software product managers/consumer product managers. Real Time-Accurate-Perfect-Structured
Use data instead of intuition: Avoid narrow decision-making perspectives and avoid unfounded decisions. Make decisions wider and more solid
The dispute between position and logic is never-ending, and data and logic chat is deadly.

2. Several stories related to data analysis

Remove the features that have an impact on the user conversion rate within half an hour;
Google’s 41 blue colors and $200 Million revenue
Can button color also make a difference in revenue?

3. Common tools for data analysis

GoogleAnalytics / MixPanel / Growing IO / Sensors Data / 友盟
Mini Program: Mini Program Data Assistant / Aladdin
Tableau / Excel / Python / Google Sheets / SQL
R / MatLab …

New understanding of Excel a few key words: Excel formulas, pivot tables, VBA, Google Script, Python Excel

4. The execution process of data analysis

Data planning (setting goals)
Data embedding/recording (technical support and realization)
Data collection and arrangement (raw data --> structured data)
Data statistics and analysis (from data --> information/knowledge)
Conclusion or action (information --> decision/action)

5. Data planning --> set goals

Reflect the operation of the product/function ☞ Select key indicators and keep an eye on them every day
Look for process bottlenecks or new product opportunities
☞ Behavior path, flow funnel
Provide product and operational decision support
☞ conduct A/B testing
☞ set the strategic water level value (when the double eleven night is a big promotion, a big red envelope will be issued to facilitate the transaction. A red envelope will be issued to the designated user based on the data)
Provide effect tracking on specific product operation strategies

4.1 How to choose key indicators

Be careful with vanity indicators (such as the promotion of a big V)
Is it related to the value of stakeholders? (Product managers look at the frequency of use of the function, investors or bosses look at GMV)
Does a better number mean a better product? (One more interface pops up for a certain function to increase the PV)
Can you communicate? (Follow changes in indicators)
Find the problem early or later? (If the frequency of user trials decreases, it means that the product has begun to decline)
Is there any information? Can you pry it? (For example, bring users back and attract users with high-quality articles)

5. Data burying point

Generally speaking, when it comes to data embedding, it refers to the collection of behavioral data;
It is essentially a record request with various clues, and it is stateless;
Data burying requirements, instead of describing the burying plan, it is better to tell BI or the engineer what you want to do;
Most of the data burying requirements are based on events (page visits are also considered events), just list the desired time list and classification;
It's difficult to bury all points at once, start with small things, such as a button, and then gradually study the bigger story;
If the company has complete burying points and data collection procedures and specifications, the above will not be counted.

6. Data collection and collation

Organize the original heterogeneous data into structured data that can be filtered/stated/processed;
Most of the work here is done by data engineers/data tools, and the product manager is not involved much in this step;

6.1 How to collect competitor data? How to collect industry data?

Use QuestMobile
Registered user id
Job Description
Competitor employees come to interview

report:

7. Data Statistics and Analysis

Indicator: The measurement value of a certain thing, such as the number of users, the number of visits, the length of the visit, the conversion rate, etc., are usually directly related to the intermediate or final goal of a specific business.
Dimensions: Different ways to subdivide indicators. For example, the number of users can be divided into new users, old users, Android users, iOS users, paid users, free users, etc. Different subdivision methods represent different analysis angles.

8. Conclusion or action

Data analysis should point to a goal, information/knowledge/decision/principle;
[Target] is the starting point and the end point of data analysis;
External: data reports, data insights, business strategy conclusions;
Internally: industry cognition, experience and principles of doing things;

9. Ideas and framework of data analysis

The preference is based on users and behaviors: who the users are, where they come from, and where they go;
Favored flow: flow cost, realization efficiency, sustainability;
Prefer to sell goods: flow, conversion, gross profit;
Favor Saas: customer classification, retention and activation, revenue scale and efficiency;
Platform-based products: the game of benefits of each role, monetization rate;

10. Data Analysis Object

User attribute data: a set of portraits;
User behavior data: a map;
Business data: a bunch of icons;
Financial data: money;
Industry data: a few figures;
Macro data

11. Common indicators and interpretations of data analysis

User attribute data-technical parameters, geographic location, age, gender, region...
User behavior data-PV, UV, (PV/UV), VV, UPV, DAU, MAU, (DAU/MAU), WAU, AAC, MAC, WAC, CTR, retention, source, access duration...
Business data-(related to you) such as: number of courses, number of training camps, number of orders, number of posts, number of packages...
Financial data-GMV, ARPU, LTV, customer unit price, repurchase rate, conversion rate, Take Rate...
Industry data-TAM, CAC, TAC...
Macro data-industry time scale, GDP share.

GMV includes data on orders placed but not paid, and orders placed but returned are also counted. Industry practice.

12. Common indicators and interpretations of data analysis

The most important thing is to have your own indicator definition, not an industry-wide [vanity indicator].
All the indicators should form a [big picture of data] in your mind to know the relationship between them;
Let data be your own language, use data to reconstruct intuition instead of emotions and likes and dislikes to construct intuition;

13. Assignments that are not homework

PDD’s financial report , comparing JD and Ali, focusing on AAC, MAC, MAU, GMV, Take Rate.
Financial report view broker, sec --> https://en.wikipedia.org/wiki/US_Securities_and_Exchange_Commission
https://sec.report/

Geek University Product Manager Training Camp: Summary of Lesson 16 of Data Analysis